Solvent and temperature effects on diastereofacial selectivity: Amines as co-solvents in n-butyllithium addition to α-chiral aldehydes

Article abstract of DOI:10.1002/ejoc.200200579

Temperature-dependent selectivity in nucleophilic additions to α-chiral aldehydes is affected by solvent mixtures. In this context, we investigated the effect on diastereoselectivity of tertiary amines as co-solvents. Addition of 5 mol% of an amine to hexane strongly influences the anti/syn ratio in the additions of nBuLi to 2-phenylpropanal and to (2S)-2-[(tertbutyl)dimethylsilyloxy]-2-propanal. In these hexane/amine mixtures as reaction solvents, we observed diastereoselectivities higher than those obtained in hexane. In the corresponding Eyring plots of ln(anti/syn) against the reciprocal temperature, we observed inversion temperatures (T_inv), these generally occurring at higher temperatures in the hexane/amine mixtures than in pure hexane. The presence of the inversion temperatures confirms that these tertiary amines take part in the solvation of the starting aldehyde, thus contributing to a solvent effect on facial selectivity. In some cases, the differential entropy of activation is the predominant factor in differentiating the diastereoselectivity among these hexane/amine mixtures. ( Wiley-VCH Verlag GmbH & Co. KGaA, 69451 Weinheim, Germany, 2003).

Full text of DOI:10.1002/ejoc.200200579

FULL PAPER
Solvent and Temperature Effects on Diastereofacial Selectivity: Amines as
Co-Solvents in n-Butyllithium Addition to α-Chiral Aldehydes
[a]
[a]
[a]
[a]
Gianfranco Cainelli,* Daria Giacomini,* Paola Galletti, and Arianna Quintavalla
Keywords: Aldehydes / Amines / Entropy / Inversion temperature / Solvent effects / Solvent mixtures
Temperature-dependent selectivity in nucleophilic additions
to α-chiral aldehydes is affected by solvent mixtures. In this
context, we investigated the effect on diastereoselectivity of
tertiary amines as co-solvents. Addition of 5 mol% of an
amine to hexane strongly influences the anti/syn ratio in the
additions of nBuLi to 2-phenylpropanal and to (2S)-2-[(tert-
butyl)dimethylsilyloxy]-2-propanal. In these hexane/amine
these generally occurring at higher temperatures in the hex-
ane/amine mixtures than in pure hexane. The presence of
the inversion temperatures confirms that these tertiary
amines take part in the solvation of the starting aldehyde,
thus contributing to a solvent effect on facial selectivity. In
some cases, the differential entropy of activation is the pre-
dominant factor in differentiating the diastereoselectivity
mixtures as reaction solvents, we observed diastereoselectiv- among these hexane/amine mixtures.
ities higher than those obtained in hexane. In the corres-
ponding Eyring plots of ln(anti/syn) against the reciprocal
temperature, we observed inversion temperatures (Tinv),
( Wiley-VCH Verlag GmbH & Co. KGaA, 69451 Weinheim,
Germany, 2003)
Introduction
vents is to be expected. Actually, in a previous paper, we
reported solvent mixture effects on diastereofacial addition
Different factors affect facial stereoselectivity;[1] among of nBuLi to some α-chiral aldehydes and imines in ether/
others, dependence on the reaction solvent has been already hydrocarbon mixtures and in binary and ternary mixtures
recognized, and several examples of solvent-dependent face of hydrocarbons. In some cases we observed enhanced dia-
selectivity have appeared in the literature.[ Solvent effects stereoselectivities in the mixtures relative to those observed
are closely related to the nature and the extent of in the pure solvents, together with the presence of tempera-
2]
[
5]
soluteϪsolvent interactions developed locally in the micro- ture-dependent solvation effects.
environments of the solute molecules.[
3]
These
This work extends those studies, and we now report on
the influence of solvent mixtures of hexane and 5 mol % of
tertiary amines (hexane/amine ϭ 95:5 molar ratio) on the
diastereoselectivity in the addition reactions of nBuLi to 2-
phenylpropanal (1) and to (2S)-2-[(tert-butyl)dimethylsily-
loxy]-2-propanal (2). We examined triethylamine (TEA),
tri-n-butylamine (Bu N), tri-n-pentylamine (Pen N),
N,N,NЈ,NЈ-tetramethylethylenediamine (TMEDA), N-
methylpyrrolidine, diisopropylethylamine (Hünig base), and
sparteine as co-solvents, and established strong effects of
such Lewis bases on the diastereofacial selectivity
soluteϪsolvent interactions are able to differentially modu-
late the free activation energies of two different stereoiso-
mers, thus exerting stereospecific solvation control over en-
antio- and diastereoselectivity. As an example, we recently
reported that, in the addition of EtMgBr to (2S)-O-(tert-
butyldimethylsilyl)lactal in a series of ethers, soluteϪsolvent
interactions govern a reversal in the de % values of the O-
protected diols obtained.[
3
3
4]
SoluteϪsolvent interactions in mixtures should be more
complex than those in pure solvents. In pure solvents, the
composition of the microsphere of solvation of the solute
is the same as in the bulk solvent, whereas in binary mix-
tures this composition can be different. The solute can in
fact interact to different extents with the two components
of the binary mixture, and this difference should affect the
composition of the solvation microsphere. In this context,
some influence on diastereoselectivity of a mixture of sol-
(
Scheme 1). The effect of Et N and TMEDA on the dia-
3
stereoselective addition of organolithium and Grignard re-
agents to ketones has already been reported, and has been
interpreted as being due to a change in the aggregation state
[6]
of the organometallic species. Our findings suggest that
tertiary amines can indeed affect the nBuLi aggregation
state, but in addition that they take also part in the solv-
ation shell of the starting aldehyde, thus contributing to a
solvent effect on the facial selectivity.
[
a]
Dipartimento di Chimica ‘‘G. Ciamician’’, University of
Bologna
Via Selmi 2, 40126, Bologna, Italy
Fax: (internat.) ϩ 39-051/209-9456
E-mail: cainelli@ciam.unibo.it
giacomin@ciam.unibo.it
A better insight into the factors affecting the stereoselec-
tivity can be obtained from studies in which the reaction
temperature is varied. Temperature-dependent measure-
Eur. J. Org. Chem. 2003, 1993Ϫ2000
DOI: 10.1002/ejoc.200200579
 2003 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
1993

G. Cainelli, D. Giacomini, P. Galletti, A. Quintavalla
FULL PAPER
and hence different stereoselectivities. At temperatures
lower than T , one cluster is present in solution, while at
inv
temperatures higher than Tinv another, different cluster is
present; Tinv would represent the interconversion tempera-
ture between these two supramolecules (Figure 1).
This interpretation has a general significance, because it
does not depend on the particular reaction investigated and
does not imply any change either in the rate-determining
[
11]
step or in the reaction mechanism.
Results and Discussion
Scheme 1
In a previous study[5] we reported a peculiar influence of
binary solvent mixtures on inversion temperatures. In par-
ments shed light on the interplay of differential activation ticular, we observed that in the addition of nBuLi to 2-phe-
enthalpies and entropies through application of the modi- nylpropanal, the presence of 5 mol % of THF in hexane was
fied Eyring Equation (1), where S is the stereoselectivity sufficient to suppress the observation of a T . In binary
inv
and k and kЈ' are the overall rate constants for the synthesis hydrocarbon mixtures, in contrast, Tinv was still present,
of the two stereoisomers.[7]
and occurred at higher temperatures than those seen in
pure solvents.
϶
϶
In this paper we analyse binary mixtures of hexane and
tertiary amines in order to ascertain their influence on dia-
stereomeric ratios and even on dynamic solvation phenom-
ena, as evidenced by the presence of inversion temperatures.
We selected a series of trialkylamines and added these (5
mol%) to hexane, thus obtaining solvent mixtures in which
the selected aldehydes 1 or 2 were dissolved. Since nBuLi is
commercially available as hexane solution, we chose hexane
as a fixed component of the binary system in order to avoid
the presence of two different hydrocarbons. In all experi-
ments, the addition reaction was performed at constant
temperature by the introduction of a stoichiometric amount
of nBuLi (2.5  hexane solution) into previously prepared
solutions of aldehydes 1 or 2 in 5 mol % amine in hexane.
The value of T was varied over a range of ca. 120 °C. The
diastereomeric anti/syn ratios (1a/1b and 2a/2b) were deter-
mined in each experiment by GC analysis (see Table 1).
Data were analysed by least-squares fitting to Equation (1)
to obtain linear correlations. For each data set we applied
a residual analysis to evaluate the number of linear trends
and to ascertain the presence of the inversion temperature.
lnS ϭ ln (k/kЈ) ϭ Ϫ(∆∆H /RT) ϩ (∆∆S /R)
(1)
Equation (1) shows a linear correlation of selectivity with
/T, but there are many experimental results that feature
1
nonlinear behaviour. In these cases, the corresponding Eyr-
ing plot of ln(k/kЈ') vs. 1/T generally consists of two straight
lines intersecting at a point defining a temperature called
the inversion temperature (Tinv).[
8]
This Tinv constitutes a break point, giving rise to two sets
of activation parameters: one for T Ͼ Tinv and one for T Ͻ
T . In previous papers we have reported that the presence
inv
of a Tinv depends on dynamic solvation effects.[ Recently,
in addition reactions of organolithium and Grignard re-
agents to (2S)-O-(tert-butyldimethylsilyl)lactal and (2S)-O-
9]
(
tert-butyldimethylsilyl)mandelicaldehyde, we observed that
the inversion temperature is only slightly dependent on the
[
10]
nucleophile, but mainly on the aldehydeϪsolvent couple.
Our results have demonstrated that temperature-dependent
changes in solvation bring about the presence of the inver-
sion temperature and tune its value. In our interpretation,
an Eyring plot featuring a Tinv is the result of two inter-
secting linear trends produced by two different solvation
clusters. These two soluteϪsolvent clusters behave as inde-
pendent species with different thermodynamic properties,
϶
϶
The ∆∆H and ∆∆S values were obtained from the slopes
and intercepts of the linear plots (Table 2) and correspond
϶
϶
϶
϶
϶
to ∆∆H ϭ ∆H
Ϫ ∆H syn, and ∆∆S ϭ ∆S
Ϫ
anti
anti
϶
∆S
syn.
϶
It is useful to remember that a negative ∆∆H derives
from a situation in which ∆H
϶
϶
Ͼ ∆H
and that there
syn
anti
is then a lower activation barrier to the formation of the
anti isomer; that is to say, the anti isomer is preferred en-
϶
thalpically. The opposite applies for ∆∆H Ͼ 0; enthalpy
then favours the syn isomer. With regard to the entropy, if
it is assumed that an addition reaction is accompanied by
϶
a loss of activation entropy, a ∆∆S Ͻ 0 derives from
∆S anti| Ͼ |∆S϶syn|; this means that the entropic loss in-
϶
|
volved in the formation of the anti isomer is larger than
that involved the formation of the syn one, so the entropy
favours the formation of the syn stereoisomer. The opposite
Figure 1. The Tinv in an Eyring plot, as the interconversion tem-
perature between two supramolecules
1994
 2003 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.eurjoc.org
Eur. J. Org. Chem. 2003, 1993Ϫ2000

Solvent and Temperature Effects on Diastereofacial Selectivity
Table 1. anti and syn (%) in the addition of nBuLi to aldehyde 1 or 2 in hexane/amine mixtures (95:5 molar ratio)
FULL PAPER
Aldehyde
Amine
T [°C]
anti (%)
syn (%)
Aldehyde
Amine
T [°C]
anti (%)
syn (%)
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
Bu
Bu
Bu
Bu
Bu
Bu
Bu
Bu
Bu
Bu
Bu
Bu
Bu
Pen
Pen
Pen
Pen
Pen
Pen
Pen
Pen
Pen
Pen
Pen
Pen
Pen
Pen
iPr
iPr
iPr
iPr
iPr
iPr
iPr
iPr
iPr
iPr
iPr
3
3
3
3
3
3
3
3
3
3
3
3
3
N
N
N
N
N
N
N
N
N
N
N
N
N
34
23.5
11.5
Ϫ10
Ϫ20
Ϫ25
Ϫ29
Ϫ38
Ϫ44
Ϫ48.5
Ϫ56
Ϫ62
Ϫ69
41
88.9
89.1
89.1
88.9
89.1
89.9
91.3
92.1
92.2
93.3
94.2
94.9
96.5
87.8
88.4
89.1
89.3
88.8
89.3
89.5
89.5
89.8
91.1
91.8
92.3
93.3
94.1
75.4
75.9
78.9
77.9
79.0
80.6
82.2
84.2
84.6
88.3
90.7
81.1
84.8
86.3
88.1
87.8
87.2
86.5
86.6
88.6
88.7
90.1
91.9
92.7
93.7
75.8
77.7
80.3
83.0
83.4
84.0
88.3
90.1
90.6
91.8
93.2
71.7
75.0
11.1
10.9
10.9
11.1
10.9
10.1
8.7
7.9
7.8
6.7
5.8
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
sparteine
sparteine
sparteine
sparteine
sparteine
sparteine
sparteine
sparteine
sparteine
sparteine
TMEDA
TMEDA
TMEDA
TMEDA
TMEDA
TMEDA
TMEDA
TMEDA
TMEDA
TMEDA
TMEDA
TMEDA
TMEDA
xfϪ19
9.5
76.1
78.5
79.5
82.1
82.1
84.6
85.6
85.7
86.8
87.6
81.5
83.9
87.1
88.8
89.1
89.2
89.5
90.3
90.9
92.7
94.2
94.7
95.3
79.6
78.1
83.1
82.1
83.1
84.8
85.3
86.6
87.5
87.9
77.2
78.9
78.5
82.0
83.6
83.3
84.2
83.8
84.3
85.8
87.5
88.0
72.8
73.4
73.1
75.7
75.3
77.0
75.8
77.9
78.5
77.5
77.3
77.2
76.9
76.5
76.3
23.9
21.5
20.5
17.9
17.9
15.4
14.4
14.3
13.2
12.4
18.5
16.1
12.9
11.2
10.9
10.8
10.5
9.7
1
Ϫ12
Ϫ19
Ϫ30
Ϫ36
Ϫ43.5
Ϫ66
Ϫ75
49
5.1
3.5
36
19.5
11.5
3
Ϫ1
Ϫ2
3
3
3
3
3
3
3
3
3
3
3
3
3
3
N
12.2
11.6
10.9
10.7
11.2
10.7
10.5
10.5
10.2
8.9
8.2
7.7
6.7
5.9
24.6
24.1
21.1
22.1
21.0
19.4
17.8
15.8
15.4
11.7
9.3
18.9
15.2
13.7
11.9
12.2
12.8
13.5
13.4
11.4
11.3
9.9
N
N
N
N
N
N
N
N
N
N
N
N
N
17.5
5.5
Ϫ5
Ϫ17
Ϫ23.5
Ϫ30
Ϫ39
Ϫ42.5
Ϫ45
Ϫ51.5
Ϫ60.5
Ϫ64
Ϫ70
50
Ϫ12
Ϫ23.5
Ϫ43
Ϫ61
Ϫ72.5
Ϫ84
32.5
12.5
Ϫ12.5
Ϫ20
Ϫ35
Ϫ50
Ϫ55
Ϫ67
Ϫ82.5
Ϫ87
56.5
32
9.1
7.3
5.8
5.3
4.7
Bu
Bu
Bu
Bu
Bu
Bu
Bu
Bu
Bu
Bu
3
3
3
3
3
3
3
3
3
3
N
N
N
N
N
N
N
N
N
N
20.4
21.9
16.9
17.9
16.9
15.2
14.7
13.4
12.5
12.1
22.8
21.1
21.5
18.0
16.4
16.7
15.8
16.2
15.7
14.2
12.5
12.0
27.2
26.6
26.9
24.3
24.7
23.0
24.2
22.1
21.5
22.5
22.7
22.8
23.1
23.5
23.7
2
2
2
2
2
2
2
2
2
2
2
EtN
EtN
EtN
EtN
EtN
EtN
EtN
EtN
EtN
EtN
EtN
40
29.5
20
1
Ϫ13
Ϫ37
Ϫ51
Ϫ61
Ϫ75
Ϫ87
53.5
30
TEA
TEA
TEA
TEA
TEA
TEA
TEA
TEA
TEA
TEA
TEA
TEA
TMEDA
TMEDA
TMEDA
TMEDA
TMEDA
TMEDA
TMEDA
TMEDA
TMEDA
TMEDA
TMEDA
TMEDA
TMEDA
TMEDA
TMEDA
21
1
Ϫ12
Ϫ19
Ϫ35
Ϫ41
Ϫ49
Ϫ61
Ϫ71
Ϫ80.5
49.5
37
N-methylpyrrolidine
N-methylpyrrolidine
N-methylpyrrolidine
N-methylpyrrolidine
N-methylpyrrolidine
N-methylpyrrolidine
N-methylpyrrolidine
N-methylpyrrolidine
N-methylpyrrolidine
N-methylpyrrolidine
N-methylpyrrolidine
N-methylpyrrolidine
N-methylpyrrolidine
N-methylpyrrolidine
TEA
TEA
TEA
TEA
TEA
TEA
TEA
TEA
TEA
11
1
Ϫ4
Ϫ11
Ϫ19.5
Ϫ22.5
Ϫ35
Ϫ40.5
Ϫ50
Ϫ67
Ϫ77.5
Ϫ87
52
29
8
1.5
8.1
7.3
6.3
Ϫ11.5
Ϫ20
Ϫ28.5
Ϫ36
Ϫ45.5
Ϫ50
Ϫ55
Ϫ64
Ϫ70
Ϫ76.5
24.2
22.3
19.7
17.0
16.6
16.0
11.7
9.9
9.4
8.2
6.8
28.3
25.0
35.7
1
Ϫ14
Ϫ24
Ϫ28
Ϫ32
Ϫ50.5
Ϫ65
Ϫ76
Ϫ88
40
TEA
TEA
sparteine
sparteine
29
Eur. J. Org. Chem. 2003, 1993Ϫ2000
www.eurjoc.org
 2003 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
1995

G. Cainelli, D. Giacomini, P. Galletti, A. Quintavalla
FULL PAPER
Table 2. Differential activation parameters and inversion temperatures found for addition of nBuLi to aldehydes 1 and 2 in hexane/amine
mixtures (95:5 molar ratio)
Entry Aldehyde Amine
Nucleophile
Tinv [°C]
T ϾTinv
T Ͻ Tinv
‡
‡
‡
‡
∆
∆H [kcal/mol] ∆∆S [cal/mol K] ∆∆H [kcal/mol] ∆∆S [cal/mol K]
1
2
3
4
5
1
1
1
1
1
Ϫ
Bu N
3
Pen
iPr
N-methyl- nBuLi
pyrrolidine
nBuLi
nBuLi
nBuLi
nBuLi
Ϫ46
Ϫ20
Ϫ37
Ϫ47
Ϫ0.9Ϯ0.1
0.04Ϯ0.01
Ϫ0.3Ϯ0.1
Ϫ0.7Ϯ0.1
Ϫ1.8Ϯ0.2
Ϫ0.8Ϯ0.2
4.1Ϯ0.2
3.0Ϯ0.4
0.2Ϯ0.4
Ϫ2.5Ϯ0.6
Ϫ0.16Ϯ0.05
Ϫ2.3Ϯ0.2
Ϫ1.6Ϯ0.2
Ϫ1.3Ϯ0.2
0.99Ϯ0.03
2.3Ϯ0.2
Ϫ5Ϯ1
Ϫ3Ϯ1
Ϫ3Ϯ1
7.6Ϯ0.1
3
N
EtN
2
[a]
0
Ϫ19[b]
Ϫ30
Ϫ34
13
12
Ϫ
Ϫ83
Ϫ
Ϫ
0.99Ϯ0.03
Ϫ1.0Ϯ0.1
Ϫ1.6Ϯ0.1
Ϫ2.8Ϯ0.1
Ϫ2.4Ϯ0.3
Ϫ1.18Ϯ0.04
0.12Ϯ0.02
Ϫ0.61Ϯ0.03
Ϫ0.7Ϯ0.1
Ϫ0.5Ϯ0.1
7.6Ϯ0.1
Ϫ0.8Ϯ0.2
Ϫ3.0Ϯ0.3
Ϫ5.9Ϯ0.3
Ϫ4Ϯ1
Ϫ0.7Ϯ0.2
3.2Ϯ0.1
0.7Ϯ0.2
0.4Ϯ0.2
0.3Ϯ0.2
Ϫ1.17Ϯ0.05
Ϫ0.9Ϯ0.1
Ϫ0.41Ϯ0.04
Ϫ1.13Ϯ0.04
Ϫ1.1Ϯ0.2
Ϫ0.9Ϯ0.2
0.3Ϯ0.4
1.8Ϯ0.2
0.1Ϯ0.2
0.3Ϯ0.7
6
7
8
9
0
1
2
3
4
1
1
1
1
1
2
2
2
2
TEA
nBuLi
nBuLi
nBuLi
sparteine
TMEDA
TMEDA
Ϫ
Ϫ
Bu
nBuLiϪTMEDA[
nBuLiϪTMEDA[
nBuLi
c]
c]
1
1
1
1
1
1.1Ϯ0.2
8.4Ϯ1.1
3
N
nBuLi
nBuLi
nBuLi
TEA
TMEDA
Ϫ35
0.26Ϯ0.04
3.6Ϯ0.2
[
a]
inv,1. [b]
T
inv,2. [c] Preformed complex.
T
holds for ∆∆S϶ Ͼ 0; the entropy now favours the anti iso- show the same ∆∆H , this means that, once again, only the
϶
϶
mer.
We first explored the addition of nBuLi to 2-phenylpro-
∆∆S plays a role in differentiating the anti/syn ratio.
We and other research groups have already recognized
panal in hexane with different amines as co-solvents. The
plots in Figure 2 refer to nBuLi addition to 1 in mixtures
of hexane with 5 mol% of tri-n-butylamine, tri-n-pentylam-
ine or diisopropyl(ethyl)amine. All three mixtures gave Eyr-
ing plots different from that obtained in pure hexane, with
increased anti predominance, especially at low temperatures
the importance of entropy in controlling stereoselective pro-
[12]
cesses.
Entropy differences are really a determinant fac-
tor, especially in enantio- or diastereoselective reactions in
which there are small differences in activation enthalpies
between the two diastereomeric transition states. We have
϶
reported some examples in which fine-tuning of ∆∆S
with Bu N and Pen N. In the high-temperature region,
3
3
could be achieved through solvent effects, so that different
reaction solvents are able to drive the diastereofacial selec-
tivity just by means of an entropy control over the domi-
Bu N and Pen N show superimposed and flattened trends
3
3
where the diastereoselectivity is constant with the tempera-
ture. This occurs because of very low differential activation
enthalpy values (Table 2, Entries 2 and 3), so the dia-
stereofacial selectivity is only determined by the differential
entropy of activation in this T region for the two amines.
[4]
nance of one stereoisomer. Now, through the use of ter-
tiary amines as co-solvents, we observed that even a small
amount of a co-solvent is sufficient for fine modulation of
diastereofacial selectivity through an entropy effect.
It has already been reported that Lewis bases (LB), such
as ethers or tertiary amines, interact with alkyllithium com-
pounds in hydrocarbon solutions to produce solvated tetra-
[
13]
meric species R Li ·nLB. In our case the presence of ter-
4
4
tiary amines could also give rise to nBuLi/amine complexes
and these could be viewed as the new organometallic spec-
ies. These complexes react with the aldehyde with their own
thermodynamic parameters and thus with a diastereoselec-
tivity different from that observed for nBuLi in pure hexane,
as we actually observed. Inversion temperatures are clearly
present in all three plots (Figure 2), and in two cases they
have different values depending on whether Bu N or Pen N
3
3
had been added (Table 2, Entries 2Ϫ4). It follows that Bu N
3
and Pen N form different soluteϪsolvent clusters with the
3
Figure 2. Eyring plots of the diastereomeric ratios obtained in the
aldehyde at low T, and these interconvert into other
solvation clusters presenting different activation parameters
and different Tinv values. The inversion temperatures in
these hexane/amine mixtures occur at higher values than in
addition of nBuLi to 1 in pure hexane (solid triangles) and in a
9
2 3
5:5 molar ratio of hexane with iPr EtN (solid circles), Pen N
(
3
solid squares), or Bu N (solid diamonds)
At low T, all these amines show fairly parallel trends. In pure hexane, as we had previously observed when studying
particular, tri-n-pentylamine and diisopropyl(ethyl)amine the addition of nBuLi to 1 in hydrocarbon mixtures.[ This
5]
1996
 2003 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.eurjoc.org Eur. J. Org. Chem. 2003, 1993Ϫ2000

Solvent and Temperature Effects on Diastereofacial Selectivity
FULL PAPER
suggests that cluster interconversion in mixtures requires
more energy, as if it were more difficult to reorganize the
solvation shell, implying the presence of a soluteϪsolvent
cluster thermodynamically more stable than that occurring
in pure solvents.
At high T, Hünig’s base presents a linear trend and a Tinv
similar to that in pure hexane. It could be surmised that
this hindered amine probably does not coordinate nBuLi,
[
14]
leaving the hexameric, homosolvated nBuLi unchanged.
The presence of a Tinv similar to that found in pure hexane
also suggests that this amine does not take part in the solv-
ation of the aldehyde, which shows the same cluster in-
terconversion behaviour as in pure hexane.
The plots in Figure 3 refer to the use of Et N and N-
3
methylpyrrolidine as co-solvents (5 mol%) in hexane.
Again, in both cases we observe an increase in the anti/syn
ratio with respect to that seen in pure hexane. Quite pecul-
iar is the behaviour of Et N, because its Eyring plot at Ϫ30
3
°
C shows an abrupt jump in selectivity. Unlike in the other
temperature ranges, from Ϫ32 to Ϫ28 °C we encountered
difficulties in collecting any data because of high scat-
[
15]
tering. This interruption marks two linear trends, one at
a higher and one at a lower T, with quite similar slopes but
different intercepts. This jump derives exclusively from a
϶
change in ∆∆S , which actually has small values but op-
posite signs in the two T regions (Table 2, Entry 6); once
again the entropy controls the diastereoselectivity of the re-
action through dynamic solvation effects. This break in the Figure 3. Eyring plots of the diastereomeric ratios obtained in the
addition of nBuLi to 1 in pure hexane (solid triangles) and in a
Eyring plot can be readily explained in terms of intercon-
9
3
5:5 molar ratio of hexane with Et N (solid diamonds) or N-meth-
version of soluteϪsolvent clusters such as happens at the ylpyrrolidine (solid circles)
inversion temperature and so, in our opinion, it can really
be considered a T . The T obtained with this amine dif-
inv
inv
fers substantially from that found in pure hexane, and this
amine and (Ϫ)-sparteine.[18] Figure 4 shows the correspond-
ing Eyring plots. Again, we obtained an increase in the pro-
portion of the anti isomer relative to when pure hexane was
used, especially with TMEDA. The two diamines display
different inversion temperatures, both of which, once again,
occur at higher values than found in hexane (Table 2, En-
confirms the presence of Et N in the solvation clusters.
3
The use of 5 mol % of N-methylpyrrolidine brings about
the presence of two inversion temperatures in the Eyring
plot (Figure 3). In our opinion, the presence of two Tinv
values can be explained by the existence of three different
solvation clusters, depending on the temperature.[
16]
From
tries 7 and 8). According to our hypothesis on T , this
inv
Ϫ87 to Ϫ19 °C there is one soluteϪsolvent cluster with its
own thermodynamic properties and its own stereoselectiv-
ity. This cluster undergoes a reorganization at Ϫ19 °C into
a second one, with a different stereoselectivity.[ However,
this second soluteϪsolvent cluster itself proved to be a
short-lived species, because at 0 °C it reorganizes into a
third, characteristic of the high T region. It is interesting to
note that the first and the last soluteϪsolvent clusters show
17]
϶
quite similar ∆∆H values, the corresponding linear trends
differing exclusively in the ∆∆S϶ values, as in the case of
Et N. In the case of this N-methylpyrrolidine/hexane mix-
3
ture as well, the different Tinv with respect to pure hexane
corroborates the presence of this amine in solvation clus-
ters. TEA and N-methylpyrrolidine mixtures show Tinv
values higher than that seen in pure hexane.
Having obtained these good results in enhancing dia-
stereofacial selectivity through the use of tertiary amines,
we therefore tested two diamines frequently used in the
Figure 4. Eyring plots of the diastereomeric ratios obtained in the
addition of nBuLi to 1 in pure hexane (solid triangles) and in a
95:5 molar ratio of hexane with (Ϫ)-sparteine (solid circles), or
chemistry of organolithium reagents: tetramethylethylenedi- TMEDA (solid diamonds)
Eur. J. Org. Chem. 2003, 1993Ϫ2000
www.eurjoc.org
 2003 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
1997

G. Cainelli, D. Giacomini, P. Galletti, A. Quintavalla
FULL PAPER
implies that the two amines enter in the solvation micros- Bu N with increasing anti/syn ratio (with respect to pure
3
phere of the aldehyde to generate different soluteϪsolvent hexane) at low T. The data obtained with these two amines
clusters with peculiar temperature-dependent interconver- do not allow the determination of a Tinv in the temperature
sion phenomena.
range explored. The absence of an inversion temperature
Alkyllithium compounds are known to be present in could be interpreted in terms of the presence of only one
solution as aggregates (RLi) , and the aggregate size de- soluteϪsolvent cluster in that temperature range. TMEDA,
n
pends on the alkyl group, solvent, concentration, and tem- meanwhile, produces a small decrease in the diastereomeric
[
19]
perature.
In hydrocarbons in particular, TMEDA has ratio relative to that seen in pure hexane, showing a Tinv
been shown to convert hexameric nBuLi aggregates into di- with a higher value than obtained in pure hexane. The dif-
[
20]
meric ones. In order to ascertain whether the presence of ference in the anti/syn ratios due to TMEDA vs. TEA or
a Tinv might be due to aggregation/disaggregation of nBuLi/ Bu N is governed by the differential activation entropy, es-
3
amine complexes, and not to dynamic solvation phenom- pecially at higher temperatures (Table 2, Entries 12Ϫ14).
ena, we studied the anti/syn ratios in the addition of the With TMEDA as co-solvent, peculiar thermodynamic par-
preformed nBuLi/TMEDA complex to phenylpropanal in ameters and a Tinv confirm the presence of this amine in
pure hexane and in a hexane/TMEDA mixture (95:5 molar the aldehyde solvation and in the coordination of lithium
ratio; Table 2, Entries 9 and 10).
as a nBuLi/TMEDA complex.
The corresponding Eyring plots are shown in Figure 5,
with comparison with those determined for pure hexane
and in the hexane/TMEDA mixture. The preformed nBuLi/
TMEDA complex, when added to phenylpropanal diluted
in hexane/TMEDA (5 mol %), shows an Eyring plot quite
similar to that obtained for the addition of nBuLi to 1 di-
luted in the hexane/TMEDA mixture (95:5 molar ratio),
similarities being seen in slopes and intercepts, but es-
pecially in T , whereas addition of the preformed nBuLi/
inv
TMEDA complex to 1 in pure hexane shows different
thermodynamic parameters and no Tinv.
Figure 6. Eyring plots of the diastereomeric ratios obtained in the
addition of nBuLi to 2 in pure hexane (solid squares), and in a 95:5
3
molar ratio of hexane with Et N (solid triangles), TMEDA (solid
3
circles), or Bu N (solid diamonds)
Figure 7 summarizes the results for the diastereomeric ex-
cess obtained for nBuLi addition to phenylpropanal (1)
with all hexane/amine 95:5 molar ratio mixtures at low and
high temperatures and compares them to data obtained in
[
21]
pure hexane and in pure THF.
The diagram clearly
Figure 5. Eyring plots of the diastereomeric ratios obtained in the
shows how the use of an amine as a co-solvent with a hy-
drocarbon is able to improve de % values meaningfully,
addition to 1 of nBuLi in pure hexane (solid triangles), nBuLi in a
9
5:5 mol mixture of hexane/TMEDA (solid diamonds), preformed
nBuLi/TMEDA complex in pure hexane (solid squares), and pre- whether at low or high T, with respect to both hexane and
formed nBuLi/TMEDA complex in a 95:5 mol mixture of hexane/
THF.
TMEDA (solid circles)
The possibility of improving the diastereofacial selectivity
through the use of a tertiary amine as a co-solvent with
These results confirm the dependence of the inversion
temperature on a phenomenon due mainly to the solvation
of the starting aldehyde and not to the nature of the nucle-
ophile.
Diastereoselection in the addition of nBuLi in the hex-
a hydrocarbon is a quite interesting aspect, especially for
synthetic and industrial applications, since THF, a com-
monly used solvent, suffers from cost and safety concerns.
[
10]
ane/amine mixture was also studied in the case of the α- Conclusion
hydroxyaldehyde (2S)-2-[(tert-butyl)dimethylsilyloxy]-2-pro-
panal (2). Figure 6 shows Eyring plots of the 2a/2b ratio
In this paper we have shown that the presence of 5mol %
obtained with 2 on treatment with nBuLi in hexane and in of tertiary amines in hexane strongly affects the diastereofa-
the presence of 5 mol % of Bu N, TEA or TMEDA. In this cial selectivity of additions of nBuLi to 2-phenylpropanal
3
case we observed superimposed linear trends of TEA and (1) and (2S)-2-[(tert-butyl)dimethylsilyloxy]-2-propanal (2).
1998
 2003 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.eurjoc.org
Eur. J. Org. Chem. 2003, 1993Ϫ2000

Solvent and Temperature Effects on Diastereofacial Selectivity
FULL PAPER
Figure 7. Diastereomeric excesses obtained at low and high temperature in pure THF and in hexane, with comparison with those found
in hexane/amine (95:5) mixtures
˚
We have demonstrated that the addition of such trialkylam- sieves (4 A). nBuLi (commercial 2.5  solution in n-hexane) was
titrated shortly before use. 2-Phenylpropanal (1) was purchased
from Aldrich. (2S)-2-[(tert-Butyl)dimethylsilyloxy]-2-propanal (2)
ines substantially enhances the preferential formation of the
anti isomers. The mode of action of such Lewis bases is to
participate in the lithium coordination sphere and in the
solvation shell of the starting aldehyde. The presence of an
was prepared from the corresponding O-protected-α-hydroxy ester
by DIBAl-H reduction.[ The hexane/amine mixtures were pre-
22]
pared in molar ratios of 95:5.
inversion temperature in fact confirms the presence of am-
ines in soluteϪsolvent clusters that are subject to tempera- Reactions: In a typical experiment, aldehyde 1 or 2 (1 mmol) was
dissolved in an anhydrous n-hexane/amine mixture (5 mol %,
ture-dependent reorganization phenomena. The Tinv values
20 mL) under inert conditions, and the solution was cooled or
obtained with these hexane/amine mixtures occur at higher
temperatures than those seen in pure hexane, as we had
previously observed when studying the addition of nBuLi
to 2-phenylpropanal in hydrocarbon mixtures. We have
demonstrated that the difference in stereoselectivity is due
to an entropy effect. The possibility of controlling the
stereoselectivities of alkyllithium reactions through the use
warmed to the desired temperature. n-Butyllithium (1.2 mmol) was
then added. After the starting aldehyde had disappeared (GC
monitoring), the reaction mixture was quenched with a saturated
4 2 2
aqueous solution of NH Cl, extracted with CH Cl (3 ϫ 50 mL),
and dried with Na SO . Chromatography of the residue on a silica
2
4
gel column gave mixtures of alcohols 1aϪ1b and 2aϪ2b, so the
chemical yields could be calculated; these ranged from 80 to 90%.
of an amine at low co-solvent concentrations is even of gen- From GC analysis of the crude products or of their corresponding
eral interest, because it would allow the use of an aliphatic O-trimethylsilyl derivatives, the anti/syn ratios and the de % values
were obtained. The average standard deviation for the de % meas-
hydrocarbon more cost- and safety-effective than the nor-
urements was less than 1%. The alcohols obtained are known prod-
mally used ethereal solvents.
ucts and their configurations were determined by comparison with
reported data.[
23]
Experimental Section
Acknowledgments
General Remarks: All reactions were performed in flame-dried
glassware under argon. GC-MS: HP5980, capillary column HP-1
or HP-5 connected to HP5970 (70 eV). GC: Fisons G8000; column:
HP-5 MS cross linked 5% PhMeSilicone, 30 m ϫ 0.25 mm ϫ 0.25
µm, chiral column OV 1701 H. TLC: Merck 60F254. Column chro-
matography: Merck silica gel 200Ϫ300 mesh. During reactions, in
order to set and maintain temperatures in the range of Ϯ 1 °C,
We wish to thank Mr. Paolo Angelucci for experimental assistance.
This work was supported by the MURST (60% and COFIN) and
the University of Bologna (fund for selected topics).
[1]
E. L. Eliel, S. H. Wilen, L. N. Mander, in Stereochemistry of
liquid N
2
/acetone baths in Dewar containers or oil baths with
Organic Compounds, John Wiley & Sons, New York, 1994, p.
water-cooling were used. Temperatures refer to the interior of the
reaction apparatus.
875.
[2]
[2a]
For recent examples see for instance:
S. Roland, P. Mang-
[2b]
eney, Eur. J. Org. Chem. 2000, 611.
G. Cainelli, D. Giacom-
ini, P. Galletti, Eur. J. Org. Chem. 1999, 61. [ M. T. Crimmins,
2c]
Starting Materials: All amines were distilled from KOH, and hex-
ane was dried by distillation from sodium and stored on molecular
A. L. Choy, J. Am. Chem. Soc. 1997, 119, 10237. [ R. W.
2d]
Eur. J. Org. Chem. 2003, 1993Ϫ2000
www.eurjoc.org
 2003 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
1999

G. Cainelli, D. Giacomini, P. Galletti, A. Quintavalla
FULL PAPER
Murray, M. Singh, B. L. Williams, H. M. Moncrieff, Tetra-
hedron Lett. 2001, 42, 7383Ϫ7385. [12b] M. Lombardo, S. Frab-
hedron Lett. 1995, 36, 2437. [ T. Saito, M. Kawamura, J. Ni-
2e]
boni, C. Trombini, J. Org. Chem. 2001, 66, 1264.
[12c]
A. Filippi,
[
2f]
shimura, Tetrahedron Lett. 1997, 38, 3231.
P. Wipf, J.-K.
N. A. Trout, P. Brunelle, W. Adcock, T. S. Sorensen, M. Sper-
Jung, Angew. Chem. Int. Ed. Engl. 1997, 36, 764. [ G. Cainelli,
2g]
anza, J. Am. Chem. Soc. 2001, 123, 6396.
[12d]
T. Sugimura, K.
D. Giacomini, F. Perciaccante, Tetrahedron: Asymmetry 1994,
Hagiya, Y. Sato, T. Tei, A. Tai, T. Okuyama, Org. Lett. 2001,
[
2h]
[12e]
5, 1913.
S. E. Denmark, N. Nakajima, O. J.-C. Nicaise, J.
3, 37Ϫ40.
Y. Inoue, H. Ikeda, M. Kaneda, T. Sumimura,
[2i]
Am. Chem. Soc. 1994, 116, 8797.
M. T. Reetz, S. Stanchev,
S. R. L. Everitt, T. Wada, J. Am. Chem. Soc. 2000, 122, 406.
[
12f]
H. Haning, Tetrahedron 1992, 48, 6813.
R. E. Rosenberg, J. S. Vilardo, Tetrahedron Lett. 1996, 37,
[12g]
[
3]
For leading references to solvent effects, see: C. Reichardt, Sol-
vents and Solvent Effects in Organic Chemistry, 2nd ed., Wiley-
VCH, Weinheim, 1990.
2185.
B. Giese, Acc. Chem. Res. 1984, 17, 438Ϫ442.
[
[
13]
14]
H. L. Lewis, T. L. Brown, J. Am. Chem. Soc. 1970, 92, 4664.
For a recent contribution on amine-complexed dimers and
tetramers of alkyllithium, see: W. Moene, M. Vos, M. Schakel,
F. J. J. de Kanter, R. F. Schmitz, G. W. Klumpp, Chem. Eur. J.
2000, 6, 225Ϫ236.
[
[
[
4]
5]
G. Cainelli, D. Giacomini, P. Galletti, P. Orioli, Eur. J. Org.
Chem. 2001, 4509Ϫ4515.
G. Cainelli, P. Galletti, D. Giacomini, P. Orioli, M. A. Polidoro,
M. C. Righetti, Helv. Chim. Acta 2000, 83, 1951Ϫ1961.
[
15]
We always encountered difficulties in collecting data for anti
and syn (%) at the inversion temperature. Around Tinv, we ob-
served a large dispersion of data, whereas at all other selected
temperatures anti/syn ratios are highly consistent with the lin-
ear trend.
6] [6a]
A. I. Meyers, J. Slade, J. Org. Chem. 1980, 45, 2585. [6b] E.
M. Carreira, J. du Bois, J. Am. Chem. Soc. 1995, 117,
106Ϫ8125.
8
[
[
[
7] [7a]
H. Eyring, J. Phys. Chem. 1935, 3, 107. [7b] S. Glasstone, K.
[
16]
J. Laidler, H. Eyring, The Theory of Rate Processes, McGraw-
Hill, New York, 1941, chapter 4.
We have already observed Eyring plots with the presence of
two Tinv points, see ref.[ and ref.
9a]
[2b]
8]
[17]
For a review see: H. Buschmann, H.-D. Scharf, N. Hoffmann,
P. Esser, Angew. Chem. 1991, 103, 480; Angew. Chem. Int. Ed.
Engl. 1991, 30, 477.
Experimental data of ln(anti/syn) collected between Ϫ19 and
0 °C differ from the theoretical median values calculated in this
T range extrapolating the two extreme linear regions, thus giv-
ing evidence for the presence of a new different solvation spec-
ies. Moreover, we do not observe any data scattering in this re-
gion.
9] [9a]
G. Cainelli, D. Giacomini, P. Galletti, Chem. Commun.
[
9b]
1
999, 567.
Angew. Chem. 1996, 108, 3016, Angew. Chem. Int. Ed. Engl.
996, 35, 2849.
G. Cainelli, D. Giacomini, P. Galletti, A. Marini,
[
[
18]
19]
1
D. Hoffmann, D. B. Collum, J. Am. Chem. Soc. 1998, 120,
5810Ϫ5811 and references cited therein.
[
10]
11]
G. Cainelli, D. Giacomini, P. Galletti, P. Orioli, Angew. Chem.
000, 112, 533; Angew. Chem. Int. Ed. 2000, 39, 523.
Inversion temperatures have indeed been found in a large num-
2
For a review of lithium solvation see: B. L. Lucht, D. B. Col-
lum, Acc. Chem. Res. 1999, 32, 1035Ϫ1042.
[
[8]
[20]
[21]
ber of completely different reactions; see for instance ref. and:
D. B. Collum, Acc. Chem. Res. 1992, 25, 448Ϫ454.
[
11a]
J. Muzart, F. H e´ nin, S. J. Aboulhoda, Tetrahedron: Asym-
We have already reported results of de % values obtained in
metry 1997, 8, 381. [
E. N. Jacobsen, J. Am. Chem. Soc. 1994, 116, 9333.
Brunne, N. Hoffmann, H.-D. Scharf, Tetrahedron 1994, 50,
11b]
M. Palucki, P. J. Pospisil, W. Zhang,
addition of nBuLi to aldehydes 1 and 2 in pure THF, see ref.
[22] [22a]
[2b]
[11c]
J.
G. Cainelli, D. Giacomini, E. Mezzina, M. Panunzio, P.
Zarantonello, Tetrahedron Lett. 1991, 32, 2967. [
22b]
M. Hir-
[
11d]
6
1
819.
T. Göbel, K. B. Sharpless, Angew. Chem. 1993, 105,
ama, I. Nishizaki, T. Shigamoto, S. I. Ito, J. Chem. Soc., Chem.
Commun. 1986, 393.
[11e]
417; Angew. Chem. Int. Ed. Engl. 1993, 32, 1329.
J. Muz-
[
23]
[2b,9b]
[23a]
art, F. H e´ nin, J.-P. P e` te, A. M’boungou-M’Passi, Tetrahedron:
See refs.
and:
M. T. Reetz, S. Stanchev, H. Haning,
Asymmetry 1993, 4, 2531. [
11f]
Tetrahedron 1992, 48, 6813.
Ber. 1991, 124, 219.
[23b]
D. Enders, S. Nakai, Chem.
I. T o´ th, I. Guo, B. E. Hanson,
Organometallics 1993, 12, 477.
[
12] [12a]
G. Cainelli, P. Galletti, D. Giacomini, P. Orioli, Tetra-
Received October 16, 2002
2000
 2003 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.eurjoc.org
Eur. J. Org. Chem. 2003, 1993Ϫ2000